Potential file does not match AIREBO/REBO style variant:

LAMMPS LAMMPS General Discussion
,
1

Using the 2Aug23 version of lammps compiled with presets I’ve been able to run the provided airebo example using lmp -sf gpu -pk gpu 1 -in in.airebo-m or running with kokkos using mpirun lmp-kgb -k on g 1 -sf kk -pk kokkos newton on neigh full comm device -airebo-m where lmp-kgb is the local name for lmp in its build folder.

that input script is:
units metal
atom_style atomic
read_data data.airebo
replicate 17 16 2
neighbor 0.5 bin
neigh_modify delay 5 every 1
pair_style airebo/morse 3.0 1 1
pair_coeff * * CH.airebo-m C H
velocity all create 300.0 761341
fix 1 all nve
timestep 0.0005
thermo 10
run 100

I am trying to run a script ( given below ) written by Eric Hahn of an airebo foam model but the model fails at reading the pair style:
Reading airebo potential file CH.airebo-m with DATE: 2016-03-15
ERROR on proc 0: Potential file does not match AIREBO/REBO style variant: # AIREBO : # AIREBO-M of T.C. O’Connor, J.W. Andzelm, M.O. Robbins (2015)
(src/MANYBODY/pair_airebo.cpp:3384)
Last command: pair_coeff * * CH.airebo-m C H

The error indicates the potentials are an incorrect match between the older 2015 O’Conner et al the newer 2016 version. There are two suggestions in the forum on a similar issue by modification of the bench/POTENTIALS/CH.airebo

Given I have not been able to track down the format of the either style variant, the problem is confusing in that the the script is in the same folder as the airebo.in script, and the pair_style and pair_coeff are identical. There is no data folder to read. The atom types is should be ok, ( just C,H) the units and atom style are the same in both cases. There is not much else that I know to look for.

Any suggestions would be appreciated.

pb

===============================================

Created by Eric N. Hahn

Simple Carbon Foam

Version 0.1

units metal
atom_style atomic
variable cubelength equal 72
region box block 0 {cubelength} 0 {cubelength} 0 ${cubelength}
create_box 2 box
create_atoms 1 random 3666 12345 box

pair_style airebo/morse 3.0 1 1
pair_coeff rebo CH.airebo-n C H

mass 1 12.0107
mass 2 1.00784
velocity all create 100.0 12345
compute 3 all pe/atom
fix 2 all balance 1000 1.05 shift xyz 10 1.05
dump eve all custom 250 dump.Cfoam_sythesis_999K.* id type x y z vx vy vz c_3
minimize 1e-10 1e-10 5000 5000
fix 1 all nve
fix temp all langevin 999 999 5 12345
timestep 0.001
thermo_style custom step temp pe etotal press vol
thermo 100
run 1000000
write_data data.Cfoam_sythesis_999K
print “Job’s done”

2

The quoted script cannot be properly read because you didn’t quote it correctly. Please see the forum guidelines.

These lines make no sense and should trigger an error.
The pair_coeff line would only be correct with a pair_style hybrid command and then it refers to the “rebo” style instead of “airebo/morse”. On top of that, the name of the potential file is quoted as “CH.airebo-n” not “CH.airebo-m”. So it would be impossible to have the kind of error you are referring to. Consequently, you are reporting form an input different to the one in your message.

3

Both are mistakes from too many trials, my apologies - I know your time is too valuable to fend off such errors. “rebo” should have been deleted.

The commands were:
pair_style airebo/morse 3.0 1 1 ( 3.0 1 1 1 was also tried )
pair_coeff CH.airebo-m C H ( “pair_coeff CH.airebo” as in the bench airebo was also attempted )

Regards,
pb

4

This information does not help. The situation is simple:

  • use the file potentials/CH.airebo-m with pair_style airebo/morse [...]
  • use the file potentials/CH.airebo with pair_style airebo [...]
  • use the file potentials/CH.rebo with pair_style rebo [...]
  • any other combination should trigger an error

The checking is based on metadata that is provided in the potential files. This is to provide inconsistent usage, since the potential files contain spline tables that need to be recreated for each variant.

If you do get an error with one of the three settings I described, you need to provide a complete input deck with the exact files or re-use files (be sure to list them explicitly) in either the bench or examples folder and just provide files that you have modified.